RadixArk Emerges from Project SGLang with $400M Valuation as AI Inference Grows
RadixArk, spun out from SGLang, has reached a $400M valuation amid rising demand in the AI inference market.
The team behind SGLang, used by companies like xAI, is now part of RadixArk. SGLang started as a project in a UC Berkeley lab. Ying Sheng, who worked at xAI, is now the CEO of RadixArk. She used to be a research scientist at Databricks.
Both SGLang and RadixArk help make AI models run better and save money. This is important because running AI services can be costly. RadixArk focuses on improving inference, which is how models give answers based on data.
Other companies, like vLLM, have also become startups from open source projects. vLLM is looking to raise about $160 million and is valued around $1 billion. Many tech companies are using vLLM for their AI needs too, showing a strong interest in these tools.
RadixArk continues to work on SGLang, which will remain free, but they have begun charging for hosting services. The need for tools that make AI work more efficiently is growing, and investments in this area are being seen more often, indicating its importance.